ACPI, IBM Thinkpad T40, and the dock
My laptop is running Linux 2.4.23, built from pristine official
sources. I've enabled ACPI and disabled APM in the kernel
configuration (built into the kernel); if I boot the laptop normally,
it behaves as expected, and I can poke around at things in /proc/acpi
and check on things like the current draw on the battery and change
the processor speed. (Expected battery lifetime increases by 66% if I
change it from the fastest setting to the slowest.) Nothing happens
if I close the lid (under APM it would suspend-to-RAM), but looking at
the sourceforge site, this is expected and I'm just doomed on that
front under a 2.4 kernel.
This all changes if I put the laptop into the dock, though. I
generally get to work, dock the laptop, then boot it; this gets me a
better display resolution when gdm starts. :-) But, ACPI is totally
scrod now. /var/log/kern.log says:
Dec 9 13:03:16 everett kernel: ACPI: Subsystem revision 20031002
Dec 9 13:03:16 everett kernel: PCI: PCI BIOS revision 2.10 entry at 0xfd906, la
st bus=15
Dec 9 13:03:16 everett kernel: PCI: Using configuration type 1
Dec 9 13:03:16 everett kernel: ACPI: IRQ9 SCI: Level Trigger.
Dec 9 13:03:16 everett kernel: ACPI: Found ECDT
Dec 9 13:03:16 everett kernel: schedule_task(): keventd has not started
Dec 9 13:03:16 everett kernel: ACPI-1120: *** Error: Method execution faile
d [\_SB_.PCI0.PCI1.DOCK.CBS2.ICFG] (Node c15aae00), AE_NOT_EXIST
Dec 9 13:03:16 everett kernel: ACPI-1120: *** Error: Method execution faile
d [\_SB_.PCI0.PCI1.DOCK.CBS2.DREG] (Node c15aabc0), AE_NOT_EXIST
Dec 9 13:03:16 everett kernel: ACPI-1120: *** Error: Method execution faile
d [\_SB_.PCI0.PCI1.DOCK._REG] (Node c15aeb40), AE_NOT_EXIST
Dec 9 13:03:16 everett kernel: ACPI: Unable to initialize ACPI objects
So what this looks like to me is that the kernel ACPI driver starts
up, looks for a list of devices, and falls over when it finds the
dock. Is this mostly right? Is there a way around it?
(And if it did work, would I be able to undock the laptop under Linux
using a magic ACPI call? :-)
--
David Maze dmaze@debian.org http://people.debian.org/~dmaze/
"Theoretical politics is interesting. Politicking should be illegal."
-- Abra Mitchell
Reply to: